2.1 Design and ethical statement
This retrospective observational study was approved by the institutional review board of Seoul National University Bundang Hospital (IRB No.B-1411/274 − 113). The requirement for patient consent was waived due to the retrospective nature of this study and minimal risk to study participants. Data were extracted from electronic medical records.
2.2 Study subjects and data collection
We collected data on patients with adult MMD who had undergone MCA-STA anastomosis at Seoul National University Bundang Hospital between May 2003 and April 2014. Two hundred forty-eight hemispheres of 208 consecutive patients met the inclusion criteria; however, 20 hemispheres were subsequently excluded due to insufficient data (8), postoperative cerebral haemorrhage (6), and postoperative cerebral infarction (6), leaving 228 hemispheres in the final sample.
Patient data were extracted from electronic medical records, including patient characteristics, details of comorbidities, and laboratory findings. BP data were collected through routine clinical practice and downloaded from our institution’s clinical data warehouse for the purpose of this study.
2.3 Surgical and anaesthetic procedures
General indication for revascularization surgery included apparent cerebral ischemia or decreased regional blood flow, vascular response, and reserve in perfusion studies in our hospital.[12] Digital subtraction angiography was performed to diagnose MMD, evaluating collateral channels and identifying appropriate donor and recipient. Thereafter, perfusion status was evaluated with computed tomography (CT) or single-photon emission computed tomography (SPECT) with acetazolamide challenge, useful for evaluating resting perfusion and reserve capacity.
Regarding anaesthetic techniques, similar protocols were followed with study participants. Radial artery cannulation was performed to perioperatively monitor blood pressure. Total intravenous anaesthesia with propofol and remifentanil was administered with a target-controlled infusion pump. Intraoperative systolic BP (ISBP) was maintained above the level of the preoperative baseline value. Normocapnia and normovolemia were maintained.
After surgery, brain CT was performed to detect any postoperative infarction or haemorrhage. Patients were admitted to a neurological intensive care unit. Neurological examination was performed once patients regained consciousness. SBP was kept within 20 mmHg of the preoperative level. Patients’ volume status was maintained as euvolemic. If neurological symptoms developed after surgery, SPECT or perfusion CT was performed to evaluate changes to the cerebral perfusion status.
2.4 Brain Perfusion Diamox SPECT image analysis
Two nuclear medicine physicians (HYL, JHK) performed independent visual assessments of the SPECT images of each cerebral hemisphere. Disagreements were resolved by discussion and consensus between the physicians. Perfusion status detected on brain SPECT images was classified as “normal”, “mild decreased”, “moderate decreased”, “severe decreased”, or “defect”. Decreased perfusion was defined as cerebral perfusion on basal SPECT within a lower colour range compared to contralateral side and/or the surrounding area. Decreased cerebrovascular reserve (CVR) was defined as abnormal cerebral perfusion within a lower colour range relative to basal SPECT, after acetazolamide administration. The severity of decreased CVR was classified into “mild”, “moderate”, “severe”, and “defect”.[13] We evaluated the concordance rate between the physicians’ evaluation, which was 89.5%.
2.5 Outcome assessment
The primary outcome of interest was the occurrence of TND after surgery. Based on a previous report,[3] TND was confirmed if the patient met all of the following criteria: (1) presence of neurological deficits that did not exist before surgery; (2) TND resolved completely within 15 days of operation; (3) presence of a significant focal increase or decrease in cerebral blood flow at the anastomosis site, captured by a postoperative SPECT scan; and (4) no hematoma or acute infarction detected on brain CT or diffusion-weighted magnetic resonance imaging.
2.6 Statistical analysis
ISBP data were summarized as a mean, categorized into four groups, using quartiles, and examined for an association with potentially confounding variables. One-way ANOVA or Kruskal-Wallis test were conducted for continuous variables, and chi-square test or Fisher’s exact test was conducted for categorical variables.
According to TND, the categorized mean ISBP was compared using the chi-square test or Fisher’s exact test. Proportion of confirmed TND cases per mean ISBP category was plotted against the patients’ cerebral perfusion status. The association between TND and mean ISBP was examined by comparing odds ratios (ORs) in a logistic regression model. ORs were adjusted for confounding variables in a multivariable model. Clinically meaningful variables as well as the variables with P-value less than 0.1 in the univariate analysis were for the model. The following variables were included in the multivariable model: age, cerebral perfusion status, operation duration, operation site, history of stroke, duration of hospitalization, pre-operation mean SBP, preoperative haemoglobin level, and post-operative haemoglobin level.
In the multivariable logistic regression model, a non-linear term representing the mean ISBP was included to examine a non-linear relationship between mean ISBP and TND risk. To control for multi-collinearity, cantered mean ISBP was used. Moreover, cerebral perfusion status was included as an interaction term to verify its impact on the association between mean ISBP and TND risk.
Finally, the proportion of confirmed TND cases per ISBP category predicted by the multivariable regression model were plotted per mean ISBP according to the cerebral perfusion status, controlling for relevant confounders, whereby averages and reference values were used for continuous and categorical variables, respectively. Statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC). P-values < 0.05 were considered statistically significant.